Maersk Line
Automating procurement across a global fleet
Automation & MLOpsTransport & Logistics
Key result+90% automation rate on requisitions
Challenge
Maersk Line wanted to automate requisition processing for a global fleet — taking all available data into account to reduce cost, optimise delivery-port selection and discover new potential suppliers.
Solution
A set of data-science algorithms was implemented as a containerised solution, exposed as a single API and integrated with existing business processes, with database persistence enabling further AI-based learning.
- Multi-level sourcing. Catalogue prices, explicit rules and past supply relationships combine to automatically infer the best supply solution.
- Cost minimisation. Total requisition cost minimised by comparing actual and estimated prices against historical supply statistics.
- Supplier focus. The algorithm guides selection towards fewer suppliers with competitive offerings and full item coverage.
Results
- +90% automation rate
- Price & cost of goods optimised
- Port selection optimised
Flowtale were able to dive right into one of our most complex processes, quickly understand our business challenge and translate that into a functional algorithm in around 4 months.
Topics
- Machine Learning
- RPA
- Data Engineering